SDAC-UAS: A Sensor Data Acquisition Unmanned Aerial …...Unit RCAT Systems Industrial UAV4 Eagle Tree Systems Flight Data Recorder Pro5 Sensors Inertial sensors 1-axis, 8 g accelerometer

AIAA SciTech 2015-0987 5-9 January 2015, Kissimmee, Florida AIAA Infotech @ 53rd AIAA Aerospace Sciences Meeting

Copyright © 2015 by the authors. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

SDAC-UAS: A Sensor Data Acquisition Unmanned Aerial Systemfor Flight State Monitoring and Aerodynamic Data Collection

Or D. Dantsker∗, Andrew V. Louis†, Renato Mancuso‡, Marco Caccamo§, and Michael S. Selig¶

University of Illinois at Urbana–Champaign, Urbana, IL 61801

This paper presents a sensor data acquisition unmanned aerial system (SDAC-UAS) for flight state moni-toring and aerodynamic data collection research on small to mid-sized unmanned aerial vehicles (UAVs). TheSDAC-UAS was developed to provide the ground-based human (safety) pilot an easily discernible display of sen-sor and state data for aircraft monitoring along with providing the ability to remotely start and stop on-boardlogging. The system is composed of three elements: an unmanned aerial vehicle, the sensor data acquisitionsystem (SDAC), and ground interface. The SDAC is a low power and low weight unit that is fitted onto theUAV and acts as the sensor data distribution hub—the SDAC combines at 100 Hz a large variety of sensorstreams into a unified high-fidelity state data stream that is simultaneously: recorded for post-flight analysis,transmitted to the ground to provide telemetry, and forwarded to a separate processing unit, such as an autopi-lot. Commands are given to the SDAC either locally using a simple interface or remotely using the down-linktransceiver connected to the ground interface. The ground interface is a portable computer setup that runs acustom graphical user interface (GUI), which displays the sensor and state data and is used to transmit com-mands to the SDAC. The entire SDAC-UAS is completely fabricated from commercial-off-the-shelf (COTS)components, which reduced cost and implementation time; and it is designed such that it can be used withalmost any small to mid-sized UAVs.

Nomenclature

ADC = analog-to-digital converterARF = almost ready-to-flyCANbus = controller area network busCOTS = commercial off the shelfCG = center of gravityDOF = degree of freedomGPS = global positioning systemGUI = graphical user interfaceIMU = inertial measurement unitI/O = input/outputI2C = inter-integrated circuitLipo = lithium polymerPFD = primary flight displayPPM = pulse position modulationPWM = pulse width modulationRC = radio controlRSSI = received signal strength indicatorSPI = serial peripheral interfaceUART = universal asynchronous receiver/transmitterUAV = unmanned aerial vehicle

∗Graduate Research Fellow, Department of Aerospace Engineering, AIAA Student Member. [email protected]†Undergraduate Research Assistant, Department of Computer Science. [email protected]‡Ph.D. Candidate, Department of Computer Science. [email protected]§Professor, Department of Computer Science. [email protected]¶Associate Professor, Department of Aerospace Engineering, AIAA Associate Fellow. [email protected]

1 of 16

American Institute of Aeronautics and Astronautics

I. Introduction

This paper presents a sensor data acquisition unmanned aerial system (SDAC-UAS) for flight state monitoringand aerodynamic data collection research on small to mid-sized unmanned aerial vehicles (UAVs). The SDAC-UASwas developed to provide the ground-based human (safety) pilot an easily discernible display of sensor and state datafor aircraft monitoring along with providing the ability to remotely start and stop on-board logging. The SDAC-UASis composed of three elements: an unmanned aerial vehicle, the sensor data acquisition system (SDAC), and groundinterface. The entire SDAC-UAS is completely fabricated from commercial-off-the-shelf (COTS) components, whichreduced cost and implementation time; and it is designed such that it can be used with almost any small to mid-sizedUAVs. A test implementation of the SDAC-UAS can be seen in Fig. 1.

At the heart of the sensor data acquisition unmanned aerial system is the sensor data acquisition system (SDAC),1–3

which is installed onto the UAV. It is both low weight and low power, operates at 100 Hz and features: a high-frequency,high-resolution six degree-of-freedom (6-DOF) inertial measurement unit (IMU) with a global positioning system(GPS) receiver, a 3-axis magnetometer, a pitot probe, an electronic tachometer, seven 10-bit analog-to-digital converters(ADC), thirty-two 12-bit analog-to-digital converters, a 14-bit analog-to-digital converter, twenty digital input/outputs(I/O), twelve pulse width modulation (PWM) signal inputs, a 40 mi downlink transceiver, an open serial, an openCANbus port, and up to 64 GB of onboard storage. The SDAC is fitted onto the UAV and acts as the sensor datadistribution hub—the SDAC combines the large variety of sensor streams into a unified high-fidelity state data streamthat is simultaneously: recorded for post-flight analysis, transmitted to the ground to provide telemetry, and forwardedto a separate processing unit, such as an autopilot. Commands are given to the SDAC either locally using a simpleinterface or remotely using the down-link transceiver connected to the ground interface.

Figure 1. A photograph of the entire system used in the test implementation of the SDAC-UAS: the aircraft instrumented with an SDAC(internal - not visible) and ground interface.

2 of 16


The ground interface is a portable computer setup that has a bi-directional data link. The computer interfaces withthe bi-directional data link over a serial connection to receive data and transmit commands. A custom graphical userinterface (GUI) runs on the computer and displays sensor and state data and is used to transmit commands to the SDAC.The GUI has 5 sub-displays that visually show: the physical state of the aircraft, the control inputs, the location of theaircraft; a primary flight display; and a raw input data feed. With the displays, the GUI has input buttons to start andstop onboard logging and make adjustments to the sub-displays. The GUI was implemented such that all the aircraftspecific data is input into a configuration file, thereby not requiring any modifications to the code to go from aircraft toaircraft.

This paper presents an implementation of the SDAC-UAS on a fixed-wing UAV. First, the paper provides backgroundand motivation behind the development of the SDAC-UAS, including similar work from other research institutions andwill also present the goals set for the development of the system. Then the next part of the paper provides details of theimplementation including descriptions of air and ground facilities. Descriptions of air facilities include information onthe airframe and instrumentation while descriptions of the ground facilities will include details of the ground interfaceand graphical user interface. The paper concludes with a summary and description of future work.

II. Background and Motivation

Situational awareness is a must for a pilot to maintain safe control of an aircraft. This becomes especially true in thecase of a ground-based human (safety) pilot flying an unmanned aerial vehicle. The pilot must see the aircraft, judge itsorientation, and detect if anything on the aircraft is faulty, all while flying the vehicle. The difficulty to do so increasesas the aircraft and/or flight path complexity increase. Often there is the added challenge of relinquishing control to anautonomous on-board system and recovering control when needed, either planned or when the aircraft flies itself into adangerous situation. Sometimes the pilot may need to fly the aircraft in a holding pattern or harder yet, following apath to way points. The difficulties in all these situations stem from the fundamental fact that the pilot is located on theground, outside of the aircraft, and must therefore visualize what the aircraft is doing.

Observing the motion of an aircraft is often not enough. Due to the increasingly complex nature of unmannedaircraft, monitoring systems must be integrated into an aircraft in order to gain a better view of the aircraft sub-systems and if all are functioning properly. There are a handful of commercially made datalogging telemetry systemsavailable,4, 5 however, they do not really offer the level of customization needed to monitor complex unmanned aircraftand their subsystems. Similarly, there are many autopilots, both open6–8 and closed-source,9–11 that offer down-link andtherefore monitoring capabilities, yet similar limitations still exist. To this end, many research institutions and agencieshave developed their own data logging and down-link systems,12–17 which are sometimes based on existing systems.Specifications for cited systems are found in Tables 1-4. By comparing the systems available, both commercial anddeveloped, the best option in terms of feature flexibility and cost seems to come from developing one’s own monitoringsystem; there will, however, be a penalty in terms of time.

Beyond the shear specifications of a monitoring system, usage requirements must also be considered. A ground-based human (safety) pilot can only glance away from the aircraft for a short period of time. The commercially madedatalogging telemetry systems mentioned4, 5 do not require any additional personnel to operate, rather they display all theinformation downlinked on a laptop screen in a mostly predefined way. The open- and closed-source autopilot systemsmentioned6–11 provide output in a similar way, however, as the name implies, open-source autopilot system groundinterfaces can be modified. The display formats for all of the systems thus far mentioned do not lend themselves to thepilot using them during flight. Custom avionic solutions12–14, 17 have similar usage requirements. Some groups that usecustom solutions even describe how a large team is required to operate and monitor the instrumented aircraft.15, 16, 18 Thegoal in developing a monitoring solution would be to make the system easy for the pilot use, specifically allowing thepilot to glance at a screen and know the state of the aircraft along with having the system alert the pilot of any ongoingor potential failures. Additionally, some form of feedback, to command the instrumentation, would be beneficial.

In order to minimize development time, the sensor data acquisition unmanned aerial system (SDAC-UAS) was to bedirectly based on the previously developed sensor data acquisition system (SDAC).1–3 The SDAC, which had performedvery well in testing and use, provided a good starting point in both hardware and software development in terms of theonboard system. The SDAC was however lacking proper integration into the rest of the aircraft and a ground interface.There were two important additions that needed to be made beyond the original SDAC. The first addition was theinclusion of motor pulse counting tachometers to measure the rotation rate of electric motors used. The second was notplacing additional sensors boards, such as the ADC boards and the newly added tachometer(s), on the SDAC board andusing the I2C protocol to transmit data, thereby minimizing the number of wire leads running from aircraft componentsto the SDAC board. An example would be wiring all the potentiometers in wing panel, which measure the control

3 of 16


surface deflection, to an ADC board and running a single I2C lead to the SDAC board; if motors were to be placedon that wing panel, the tachometers could also be connected to that same I2C lead, which would highly simplify theproblem of wiring a wing, especially in the case of a tiltrotor.19, 20 The ground half of the SDAC-UAS system, whichwould be developed from scratch, would need to be modular, quick to set up, and provide easy viewing of importantinformation in any light setting.

Table 1. Commercially-made datalogging telemetry units

Unit RCAT Systems Industrial UAV4 Eagle Tree Systems Flight Data Recorder Pro5

Sensors

Inertial sensors 1-axis, ±8 g accelerometer 2-axis, ±38 g accelerometerMagnetometers - -Altimeter (barometric) 8 ft resolution 1 ft resolutionAirspeed (pitot probe) 10–290 mph 9–350 mphGPS 1 Hz 10 HzDigital I/O - -Analog inputs 2 2Other inputs 2 Thermocouples, current and voltage measurement, optical

RPM measurement2 Thermocouples, current and voltage measurement, opticalRPM measurement, 4 CH PWM measurement

Data Handling

Sampling rate 20 Hz 40 HzLocal output - -Storage up to 512 MB SD 10 kB on-boardRF link 15 mi 14 mi

Estimated cost $2,500+ $650–1,500+

Table 2. Open-source commercially-made autopilots

Unit Paparazzi Lisa/M6 3D Robotics APM 2.67 Pixhawk PX4 Autopilot8

Sensors

Inertial sensors 3-axis, ±2-16 g accelerometer 3-axis,±250-2000 deg/s gyroscope

3-axis, ±2-16 g accelerometer 3-axis,±250-2000 deg/s gyroscope

3-axis, ±2-16 g accelerometer 3-axis,±245-2000 deg/s gyroscope

Magnetometers 3-axis ±8 G 3-axis ±8 G 3-axis ±2-12 GAltimeter (barometric) 1 ft resolution 1 ft resolution 0.3 ft resolutionAirspeed (pitot probe) Add-on supported Add-on supported 0–223 mphGPS 5 Hz 5 Hz 5 HzDigital I/O 3 0-12 0Analog inputs 7 0-12 (same pins as Dig I/O) 2Other inputs 1x CANbus, 1x SPI, 1x I2C 8 PWM signals, 1x I2C, 2x serial Up to: 1x PPM sum, 1x RSSI, 6x UART,

2x SPI, 3x I2C, and 1x CANbus

Data Handling

Sampling rate 50 Hz 50 Hz 50 HzLocal output Serial Serial SerialStorage 512 KB on-board 16 MB on-board 2 MB on-boardRF link Add-on supported Short 15 mi

Estimated cost $210 $240+ $200

4 of 16


Table 3. Closed-source commercially-made autopilots

Unit Cloud Cap Piccolo II9 MicroPilot MP2128g10 Kestrel Autopilot v2.411

Sensors

Inertial sensors 3-axis, ±10 g accelerometer 3-axis,±300 deg/s gyroscope

3-axis, ±5 g accelerometer 3-axis gyro-scope

3-axis, ±10 g accelerometer 3-axis,±300 deg/s gyroscope

Magnetometers Add-on supported Add-on supported 2-axis and 3-axisAltimeter (barometric) 1 ft resolution 1 ft resolution 0.8 ft resolutionAirspeed (pitot probe) up to 180 mph up to 300 mph 0–130 mphGPS 4 Hz 4 Hz 4 HzDigital I/O 16 8 12Analog inputs 4x 10 bit 32x 24 bit at 5 Hz 3x 12 bitOther inputs CANbus - 4-8 PWM signals, 4 Serial Ports (Std,

SPI, I2C)

Data Handling

Sampling rate 20 Hz 5–30 Hz 100 HzLocal output LPT Serial SerialStorage - 1.5 MB on-board 512 KB on-boardRF link 25 mi 3 mi 15 mi

Estimated cost $20,000+ $6,000+ $2,500+

Table 4. Custom avionic system solutions

Unit Higashino and Sakurai12 Beard et al13 FCS-2014

Sensors




Magnetometers - - -Altimeter (barometric) - (yes) (yes)Airspeed (pitot probe) (5-hole) (yes) (yes)GPS - 1 Hz 4 HzDigital I/O - - 12Analog inputs 16x 12 bit 16x 12 bit 2x 16 bit, 8x 16 bitOther inputs - 4x serial 4x RS-232

Data Handling

Sampling rate 20Hz 130 Hz 100 HzLocal output - - -Storage 12 MB on-board up to 512 KB on-board 64 MB on-boardRF link Serial communication line 3 mi Supported

Unit NASA EAV15 NASA AirSTAR16 Brusov et al. PRP-J517

Sensors



3-axis, ±2-6 g accelerometer 3-axis,±300 deg/s gyroscope

Magnetometers - - -Altimeter (barometric) - (yes) (yes)Airspeed (pitot probe) (5-hole) (yes) (yes)GPS (yes) 5 Hz -Digital I/O - 2 0Analog inputs 16x 12 bit 48x 16 bit 24x 12 bitOther inputs 8x PWM signals, 4x RS-232, 8x serial,

1x CANbus3x serial 4x PWM signals

Data Handling

Sampling rate 10Hz 50 Hz 100 HzLocal output - - -Storage 2x 8 GB CF up to 512 KB on-board Up to 512 MB SDRF link (yes) (yes) -

5 of 16


III. Test Implementation

The SDAC-UAS was implemented on a fixed-wing trainer-type radio control model airplane in order to test thesystem. The SDAC, which had previously been installed onto the aircraft,1, 2 was upgraded and then modified tocommunicate with a newly developed ground interface. This section provides information about the tested system,particularly the air facilities, the aircraft and on-board instrumentation; and the ground facilities, the ground interfaceand graphical user interface.

A. Air Facilities

The air facilities of the SDAC-UAS are made up of the aircraft and the instrumentation integrated into the aircraft. Theaircraft is a standard COTS RC trainer airplane that is structurally unmodified. The instrumentation installed into theaircraft is based on the previously developed SDAC system and was integrated into the aircraft proving measurementsof all aircraft inputs and motion.

1. AircraftA radio control model airplane was built to test the sensor data acquisition unit.1–3 The aircraft used was a Great

Planes Avistar Elite,21 which is a 62.5 in wingspan, 7-8 lb fixed-wing trainer-type airplane. It is equipped with anelectric propulsion system that uses an AXI 4120/14 600 W motor,22 a Castle Creation Phoenix ICE 75 Amp electronicspeed controller,23 and a Thunder Power 14.8 V, 5 Ah lithium polymer (Lipo) battery.24 The model is actuated usingFutaba S3004 ball-bearing standard-torque servos and is controlled by a 2.4 GHz R6014HS spread spectrum receiver.25

The radio control system is powered by a Castle Creations CC BEC regulator, which uses a Thunder Power 7.4 V,450 mAh lithium polymer battery. The completed flight-ready aircraft is shown in Fig. 2, its physical specifications aregiven in Table 5, and its airframe component specifications are given in Table 6.

Figure 2. Flight-ready Great Planes Avistar Elite model aircraft.

6 of 16


Table 5. Great Planes Avistar Elite unmanned aircraft physical specifications

Geometric Properties

Overall Length 55.0 in (1395 mm)

Wingspan 62.5 in (1590 mm)

Wing Area 672 in2 (43.3 dm2)

Aspect Ratio 6.62

Inertial Properties

Weight

Empty (w/o Battery) 7.53 lb (3.415 kg)

4S LiPo Battery 1.17 lb (0.530 kg)

Gross Weight 8.70 lb (3.945 kg)

Wing Loading 29.8 oz/ft2 (90.9 gr/dm2)

Table 6. Great Planes Avistar Elite unmanned aircraft airframe component specifications

Construction Built-up balsa and plywood structure, aluminum wing tube, aluminum landing gear, ABS canopy, andplastic-film sheeted.

Flight Controls

Controls Ailerons (2), elevator, rudder, throttle, and flaps (2)

Transmitter Futaba T14MZ

Receiver Futaba R6014HS

Servos (8) Futaba S3004

Regulator Distribution Castle Creations CC BEC

Receiver Battery Thunder ProLite 20c 2S 7.4V 450 mAh

Propulsion

Motor AXI 4120/14 Outrunner

ESC Castle Creation Phoenix ICE 75 Amp Brushless Speed Controller

Propeller APC 13x8E

Motor Flight Pack Thunder Power ProPower 30c 4S 14.8 V 5 Ah lithium polymer battery

Flight Time 10–20 min

7 of 16


2. InstrumentationThe testbed aircraft was instrumented with an updated version of the sensor data acquisition system (SDAC),1–3

which can be seen in Fig. 3. The SDAC was developed from COTS components and is plug-and-play, meaning that itcould easily be installed into almost any aircraft. As mentioned earlier, the unit operates at 100 Hz and includes: a high-frequency, high-resolution six degree-of-freedom (6-DOF) inertial measurement unit (IMU) with a global positioningsystem (GPS) receiver, a 3-axis magnetometer, a pitot probe, an electronic tachometer, seven 10-bit analog-to-digitalconverters (ADC), thirty-two 12-bit analog-to-digital converters, a 14-bit analog-to-digital converter, twenty digitalinput/outputs (I/O), twelve pulse width modulation (PWM) signal inputs, a 40 mile downlink transceiver, an openserial, an open CANbus port, and up to 64 GB of onboard storage. Given the included sensors, the system is able tosimultaneously log and transmit: 3D linear and angular accelerations, velocities, and position along with GPS location;pitot probe airspeed; 3D magnetic field strength and heading; control surface inputs; and control surface deflections.The performance specifications for the updated SDAC are given in Table 7. A description of the software architectureused in the implementation is given in Mancuso et al.1

Figure 3. A photograph of the original sensor data acquisition system (SDAC) unit.

The updated SDAC was fitted onto the aircraft and acts as the sensor data distribution hub for the various sensorsinstalled. A system diagram depicting the specific configuration of the instrumentation, along with the flight control andpropulsion systems, is shown in Fig. 4. Starting from the top-left of the diagram, the RC receiver outputs PWM controlsignals to servos and ESC, while a duplicate stream of PWM control signals are sent to the SDAC. The receiver getsits power from a Lipo battery connected through a regulator. The ESC, which drives the motor, draws power from itsown battery. In the center of the diagram, the SDAC is connected to a variety of devices: an IMU, 3D magnetometer, 4ADCs, an RPM sensor and a telemetry radio. Three of the ADCs are connected to potentiometers to measure controlsurface deflections while the last is used to measure the voltages of each of the batteries. The SDAC acquires data fromthe sensors and from the stream of PWM control signals coming from the receiver and outputs a unified stream to thetelemetry radio while simultaneously logging it. The output stream can also be transmitted to another on-board device.The last two systems in the diagram are the telemetry radio and the video system, camera and transmitter, which wasadded to simulate what would be found on larger aircraft. Each of these systems are powered by separate Lipo batterieswith regulators to output the correct voltage. As mentioned before, one of the ADCs is being used to measure thevoltages of all the batteries. This ADC is connected to the raw output of each of the batteries through voltage dividerscircuits that scale the voltages of the batteries to the range the ADC can measure. The specifications of the componentsused in the updated, tested sensor data acquisition system are given in Table 8 and information about the installation ofthese components is described in Dantsker et al.2

8 of 16


Aircraft Systems and Components

Motor

900 MHz

Telemetry Radio

2.4 GHz

R/C Receiver

Signal (line color):

█ Flight controls

█ Sensor data

█ State data

█ Transmission

█ Power

Wiring (line type) :

— Signal and power

---- Signal only

Power (shading color):

█ Flight controls

█ Motor

█ Avionics

█ Telemetry

█ Video

…... Power only

Key

ESC

Regulator

SDAC

microSD

IMU

GPS

Pitot 3D-Mag 3x ADC

Pots

Pulse

Tachometer

Motor

LiPo

Radio

LiPo

RX

LiPo

ADC

Servos

Camera

5.8 GHz

Video TX

Video

LiPo

Regulator

Regulator AVI

LiPo

Regulator

Figure 4. A block diagram of the aircraft systems.

9of16

Am

ericanInstitute

ofAeronautics

andA

stronautics

Table 7. Updated sensor data acquisition (SDAC) system performance specifications

Sensors

Inertial sensors 3-axis, ±18 g accelerometer 3-axis, ±300 deg/s gyroscopeMagnetometers 3-axis ±750 mG and 3-axis ±11 GAltimeter (barometric) 1 ft resolutionAirspeed (pitot probe) 5–180 mphGPS position Up to 120 Hz (IMU assisted)Tachometer Up to 4 brushless motor pulse countersDigital I/O Up to 20PWM inputs Up to 12Analog inputs Up to 7x 10 bit, 32x 12 bit, 1x 14 bitFurther expansion capabilities I2C, 1x serial port, CANbus

Data Handling

Logging rate 100 HzLocal output Serial or EthernetStorage Up to 64 GB microSDRF link 40 miRF rate 10-25 Hz

Table 8. Tested sensor data acquisition (SDAC) system component specifications

Processing unit BeagleBone running 32-bit Ubuntu Linux

Sensors

IMU XSens Mti-g 6-DOF IMU with Wi-Sys WS3910 GPS Antenna

Airspeed probe EagleTree Systems pitot-static probe

Airspeed sensor All Sensors 20cmH2O-D1-4V-MINI differential pressure sensor

Magnetometer PNI Corp MicroMag 3

Analog-to-digital converters 4x Gravitech 12 bit - 8 Channel ADC

Potentiometers 6x BI Technologies 6127

Tachometer 1x Sparkfun ProMicro

Power

Regulators 3x Castle Creations CCBEC

Batteries 1x Thunder Power ProLite 2S 450 mAh (avionics)2x Thunder Power ProLite 3S 1350 mAh (telemetry and video)

Telemetry transceiver Digi 9X Tend 900-MHz card

Data Storage 8GB microSD card

10 of 16


B. Ground Facilities

The ground facilities of the SDAC-UAS are made up of a ground interface and a graphical user interface (GUI). Theground interface is a portable computer setup that has a bi-directional data link with the instrumented aircraft. The GUI,which runs on the computer, displays sensor and state data and is used to transmit commands to the SDAC. The groundfacilities are designed to provide the human pilot with a quick view of the state of the aircraft and the ability to start orstop on-board data logging.

1. Ground InterfaceThe ground interface is a portable computer setup that has a bi-directional data link to communicate with the

instrumented aircraft; the ground interface can be seen in Figure 5. The portable computer setup used for the SDAC-UASimplementation consists of two pieces: a mini-ITX tower and a tripod monitor rig with the bi-directional data link; thespecifications of all components used to aseemble the portable computer setup can be found in Table 9. The groundinterface is set up at the flying site by simply opening and expanding the tripod legs, threading the monitor-data linkstand onto the tripod, connecting the video and USB cables from the tripod rig to the tower, and finally plugging thepower cables into a powered strip or outlet. The ground interface was built with power efficient components, includinga processor with integrated graphics, a solid state hard drive, a high-efficiency power supply, and an LED lit monitor,in order to minimize the power requirements at the flying site. The total power required for the ground interface isless than 200 W, allowing it to easily be powered by a car battery inverter. In addition, the tower uses all solid statecomponents, which makes it insensitive to movement or vibrations.

Figure 5. A photograph of the sensor data acquisition system unmanned aerial system (SDAC-UAS) ground interface during flight.

11 of 16


Table 9. Tested portable computer setup component specifications

Mini-ITX Tower

Processor Intel Core i7-4790S

Memory PNY XLR8 8GB (2 x 4GB)

Hard Drive Samsung 840 EVO 250GB SSD

Motherboard ASUS H97I-PLUS

Power Supply SeaSonic S12G-450

Case Rosewill Legacy U2-B

Tripod Monitor Rig

Tripod Manfrotto 3221

Monitor Dell P2314H 23” HD Monitor with LED backlight

Telemetry transceiver Digi 9X Tend 900-MHz card with RS-232/485/422 interface

Serial-to-USB adapter Tripp Lite USA-19HS

The tripod rig was constructed as following. An adapter plate was machined to the size of the monitor stand basewith a 3/8”-16 threaded hole placed vertically beneath the center of gravity of the monitor. The threaded hole matchesthe tripod and allows the plate to be easily threaded to and un-threaded from the tripod. The adapter plate was fastenedto the base of the monitor stand. The monitor was then attached to the stand using the standard mounting procedure.Afterward, the telemetry radio, which is the same model used on the aircraft, was connected to an RS-232/485/422interface, then fastened to the back of the stand with the antenna pointing upwards. The telemetry radio was connectedto a serial-to-USB converter which was also attached to the monitor stand. A trackball mouse, used to interface withthe computer, was adhered to top of the monitor stand base. A trackball mouse was used as there is too limited of aworking area to use a regular mouse. Proper cable connection and routing was then performed. For the purpose oftesting, a stand alone keyboard was used to debug problems, however, is not required for normal operation of the groundinterface.

2. Graphical User InterfaceA custom graphical user interface (GUI) was implemented on the ground station interface to display sensor and

state data and transmit commands to the aircraft instrumentation. The GUI was designed to provide the human pilotwith a quick view of the state of the aircraft. The interface has five sub-displays that show: the physical state of theaircraft, the control inputs, the location of the aircraft; a primary flight display; and a raw input data feed. Along withthe displays, the GUI has input buttons to start and stop onboard logging and make adjustments to the sub-displays. TheGUI was implemented such that all the aircraft specific data is input into a configuration file, thereby not requiring anymodifications to the code to go from aircraft to aircraft. The graphical user interface, which was implemented in a 16:9ratio to be displayed on today’s high resolution monitors, can be seen in Figure 6.

The top left GUI sub-display shows the physical state of the aircraft. The sub-display shows position of all thecontrol surfaces, two control surfaces on each of the wings (right and left), one control surface on each of the stabilizers(right and left), and one control surface on the vertical stabilizer. The aircraft displayed has a T-tail in order to makeviewing of all the control surfaces easier. The control surfaces movement displayed are driven by potentiometermeasurements and may be scaled as desired. Control surfaces displayed may also be disabled or may be assigned thesame values of other control surfaces; e.g., the case of an aircraft with only ailerons and no flaps, the interior wingcontrol surfaces may either be held at center or moved to the same position as the exterior wing control surfaces. Theaircraft displayed also provides the pilot a visual cue if a control surface jams during flight; the GUI monitors themeasured deflection and compares it to the control input expected deflection, and if the two values do not match within aset error range, the control surface will be highlighted by a flashing red box, which is accompanied by an audible alarm.

The physical state of the aircraft sub-display also shows the rotation rate of the motor. The current implementationof the GUI is set up to allow up to four motor rotation rates to be shown, however it should be mentioned that the currentversion of the SDAC could in theory allow for up to 124 motor pulse based rotation rates units to be used simultaneously.Finally, the sub-display also shows voltage level of up to six batteries, visually using a bar and numerically below; thecolors of the voltage level bars change from green to yellow to red depending on the battery voltage and the assignedtransition points. Once a battery hits a red voltage transition point, an audible alarm will sound to alert the pilot. The

12 of 16


Figure 6. A screenshot of the GUI during flight.

13of16

Am

ericanInstitute

ofAeronautics

andA

stronautics

number of how many batteries are measured could also be expanded depending on how many of the SDAC ADCchannels are used for voltage measurement.

The bottom left GUI sub-display shows the control inputs sent to the aicraft, specifically showing the position of theRC transmitter joysticks, switches, and sliders. There are two 2-axis joysticks, ten switches, and seven sliders. Thepositions of each of the control inputs to the aircraft are shown in the absolute, meaning that any curve modificationsapplied between the physical motion of the control sticks on the transmitter to the output of the transmitter aredisregarded, only the control inputs received by the aircraft actuators are shown.

The top right GUI sub-display shows the location of the aircaft using the aircraft GPS measurements. The sub-display provides an overhead map with the view either fixed with the home point at center, so called ”Fixed Mode”, orfollowing the aircraft with the aircraft at the center, so called ”Follow Mode.” The map zoom, given in map tile units,may be changed along with the size of the aircraft icon. The map shown in the sub-display is also set up such thatcolored boundary lines can be drawn on the map to show the pilot a visual indication of when the aircraft is going to flyoutside of an intended flight zone. The boundary lines automatically adjust to changes in zoom and map motion, aswould occur in ”Follow Mode.” The GUI can also be set up to provide a visual and audible alarm when a boundary lineis crossed. Finally, the altitude of the aircraft is displayed at the bottom right of the sub-display.

The bottom middle GUI sub-display provides a primary flight display (PFD). The PFD implemented26 provides apitch ladder, a bank angle indicator, an indicated airspeed indicator, a ground speed readout, an altitude above sea levelindicator, and a heading indicator. All the values displayed in the PFD are given in metric units.

Lastly, the bottom right GUI sub-display provides the raw input data feed. The sub-display shows the full datamessages being transmitted by the SDAC and received by the ground interface. These messages are useful in order todiagnose system failures; e.g., if the the raw input data feed no longer updates, the pilot knows there is a problem withdata transmission. The messages are also used in order to set up the aircraft configuration file. The messages providethe digital values for each of the sensor outputs, which are used to set the minimum, middle, and maximum values foreach display parameter in the configuration files.

Like most modern graphical applications, the GUI uses an event-driven programming model. This means thatthere is a main event loop that queues and processes all events that are triggered by the program. The main event loopprocesses events by running the callback functions of any objects that have registered as listeners. This event-drivenprogramming model has a useful optimization of detecting the rapid fire of a single event and running the appropriatehandlers once. In an effort to take advantage of modern multi-core processors, a thread is dedicated to reading andprocessing data received from the SDAC through the wireless link. This thread is separate from the main event loop.It reads data from the serial connection as a string and then parses the data from the string. After parsing, the data isplaced in a buffer. Finally, the thread triggers an event and provides listeners with a reference to the buffer. When theGUI is initialized, the main window is registered as a listener for data sample read events. When its event handler istriggered, the main window copies the data to the proper sub-displays. After copying, a user interface repaint event istriggered. This process continues for the lifetime of the program. By separating the serial input and output from themain event-loop, the user interface remains responsive while the input thread blocks waiting for the next data sample.Dedicated graphics processing units are increasingly common in modern computers. Knowing this fact, the widgetscomposing the GUI were made to use OpenGL so that rendering can be accelerated by dedicated hardware when it ispresent.

The GUI requires that the host system have the Qt (open source) libraries27 installed, which are cross platform.The GUI has rather minimal system requirements and thus should be able to run smoothly on any computer with arecent processor and hardware accelerated graphics support. In the current implementation, the GUI ran on an IntelCore i7-4790S using just the processor integrated graphics for rendering acceleration. The GUI only utilized 2% of theprocessor resources and therefore did not stress the system.

14 of 16


IV. Summary and Future Work

This paper describes the implementation of a sensor data acquisition unmanned aerial system (SDAC-UAS) forflight state monitoring and aerodynamic data collection research on small to mid-sized unmanned aerial vehicles (UAVs).The SDAC-UAS was developed to provide the ground-based human (safety) pilot an easily discernible display of sensorand state data for aircraft monitoring along with providing him the ability to remotely start and stop on-board logging.The system is composed of three elements: an unmanned aerial vehicle, the sensor data acquisition system (SDAC), andground interface. The SDAC acts as the sensor data distribution hub that combines at 100 Hz a large variety of sensorstreams into a unified high-fidelity state data stream that is simultaneously: recorded for post-flight analysis, transmittedto provide telemetry, and forwarded to a separate processing unit. The ground interface is a portable computer setupthat runs a custom graphical user interface (GUI), which displays the sensor and state data and is used to transmitcommands to the SDAC. The entire SDAC-UAS was completely fabricated from commercial-off-the-shelf (COTS)components, which reduced cost and implementation time; and it is designed such that it can be used with almost anysmall to mid-sized UAVs.

Future work to improve the current implementation of the SDAC-UAS will include the miniaturization of theground interface, from the tripod rig and tower to a light weight, battery-powered tablet. The tablet would be placedonto the RC transmitter and would eliminate the need for most of the wiring along with the mouse. Beyond makingimprovements to SDAC-UAS ground interface, most of the components in the onboard instrumentation will also beminiaturized by transitioning from through-hole components and perfboards to surface-mount components and customprinted circuit boards. The change would decrease the amount of manual wiring and therefore ease assembly alongwith reducing the system weight.

Acknowledgments

The material presented in this paper is based upon work supported by the National Science Foundation (NSF)under grant numbers CNS-1302563 and CNS-1219064. Any opinions, findings, and conclusions or recommendationsexpressed in this publication are those of the authors and do not necessarily reflect the views of the NSF.

Figure 7. A photograph of the test implementation of the SDAC-UAS in use.

15 of 16


References1Mancuso, R., Dantsker, O. D., Caccamo, M., and Selig, M. S., “A Low-Power Architecture for High Frequency Sensor Acquisition in

Many-DOF UAVs,” International Conference on Cyber-Physical Systems, Berlin, Germany, Apr. 2014.2Dantsker, O. D., Mancuso, R., Selig, M. S., and Caccamo, M., “High-Frequency Sensor Data Acquisition System (SDAC) for Flight Control

and Aerodynamic Data Collection Research on Small to Mid-Sized UAVs,” AIAA Paper 2014-2565, AIAA Applied Aerodynamics Conference,Atlanta, Georgia, June 2014.

3Dantsker, O. D., Mancuso, R., Caccamo, M., and Selig, M. S., “Maneuver-Aware Sensor Fusion Layer for Agile Electric UAVs,” Submitted toIEEE International Conference on Cyber-Physical Systems, Seattle, Washington, Apr. 2015.

4RCAT Systems, “RCAT Systems - UAV & Unmanned Vehicle Products, UAV Telemetry System, UAV Electronics, Pitot Probes, Alpha BetaProbes,” http://rcatsystems.com/uav.php, Accessed Oct. 2013.

5Eagle Tree Systems, LLC, “Eagle Tree R/C Telemetry,” http://www.eagletreesystems.com/, Accessed Oct. 2013.6The Paparazzi Project, LLC, “Paparazzi,” http://paparazzi.enac.fr/, Accessed Nov. 2012.73D Robotics, “APM — Open source autopilot,” http://ardupilot.com/, Accessed Nov. 2013.8PX4, “PX4 Autopilot Platform,” https://pixhawk.ethz.ch/, Accessed Nov. 2013.9Cloud Cap Technology, “Cloud Cap Technology – Piccolo II highly integrated UAS Autopilot,” https://www.cloudcaptech.com/piccolo II.shtm,

Accessed Oct. 2013.10MicroPilot, “MicroPilot - Products - MP2128g,” http://www.micropilot.com/products-mp2128g.htm, Accessed Oct. 2013.11Lockheed Martin, “Kestrel Flight Systems,” http://www.lockheedmartin.com/us/products/procerus/kestrel.html, Accessed Nov. 2013.12Higashino, S. I. and Sakurai, A., “A UAV Flight-Experiment System for the Estimation of Aerodynamic Characteristics,” AIAA Paper

2003-6584, AIAA “Unmanned Unlimited” Conf. and Workshop & Exhibit, San Diego, CA, Sept. 2003.13Beard, R. W., Kingston, D., Quigley, M., Snyder, D., Christiansen, R., Johnson, W., McLain, T., and Goodrich, M. A., “Autonomous Vehicle

Technologies for Small Fixed-Wing UAVs,” Journal of Aerospace Computing, Information, and Communication, Vol. 2, No. 2, 2005, pp. 92–108.14Christophersen, H. B., Pickell, R. W., Neidhoefer, J. C., Koller, A. A., Kannan, S. K., and Johnson, E. N., “A Compact Guidance, Navigation,

and Control System for Unmanned Aerial Vehicles,” Journal of Aerospace Computing, Information, and Communication, Vol. 3, No. 2, 2006,pp. 187–213.

15Ippolito, C., Yeh, Y. H., and Kaneshige, J., “Neural Adaptive Flight Control Testing on an Unmanned Experimental Aerial Vehicle,” AIAAPaper 2007-2827, AIAA Infotech@Aerospace, Rohnert Park, CA, May 2007.

16Jordan, T. L. and Bailey, R. M., “NASA Langley’s AirSTAR Testbed: A Subscale Flight Test Capability for Flight Dynamics and ControlSystem Experiments,” AIAA Paper 2008-6660, AIAA Atmospheric Flight Mechanics Conference, Honolulu, HI, Aug. 2008.

17Brusov, V., Grzybowski, J., and Petruchik, V., “Flight Data Acquisition System for Small Unmanned Aerial Vehicle,” Proceedings of theInternational Micro Air Vehicles, ’t Harde, The Netherlands, Sept. 2011.

18Jordan, T. L. and Bailey, R. M., “AirSTAR: A Subscale Flight Test Facility For Experimental Validation,” NASA Aviation Safety TechnicalConference, St. Louis, MO, Oct. 2008.

19Rothhaar, P. M., Murphy, P. C., Bacon, B. J., Gregory, I. M., Grauer, J. A., Busan, R. C., and Croom, M. A., “NASA Langley DistributedPropulsion VTOL Tilt-Wing Aircraft Testing, Modeling, Simulation, Control, and Flight Test Development,” AIAA Paper 2014-2499, AIAA AviationTechnology, Integration, and Operations Conference, Atlanta, Georgia, June 2014.

20Busan, R. C., Rothhaar, P. M., Croom, M. A., Murphy, P. C., Grafton, S. B., and O-Neal, A. W., “Enabling Advanced Wind-Tunnel ResearchMethods Using the NASA Langley 12-Foot Low Speed Tunnel,” AIAA Paper 2014-3000, AIAA Aviation Technology, Integration, and OperationsConference, Atlanta, Georgia, June 2014.

21Hobbico, Inc., “Great Planes Avistar Elite .46 Advanced Trainer RTF,” http://www.greatplanes.com/airplanes/gpma1605.html, Accessed Oct.2013.

22Model motors s.r.o., “AXI 4120/14 GOLD LINE,” http://www.modelmotors.cz/index.php?page=61&product=4120&serie= 14&line=GOLD,Accessed Oct. 2013.

23Castle Creations, Inc., http://castlecreations.com/, Accessed Oct. 2013.24Advanced Energy Tech, “Thunder Power RC,” http://thunderpowerrc.com/, Accessed Nov. 2012.25Hobbico, Inc., “Futaba Radio Control Systems and Accessories,” http://futaba-rc.com/, Accessed Oct. 2013.26Marek Cel, “QFlightInstruments,” http://sourceforge.net/projects/qfi/, Accessed May. 2014.27“Qt Project,” http://qt-project.org/, Accessed May. 2014.

16 of 16


Documents

SDAC-UAS: A Sensor Data Acquisition Unmanned Aerial …...Unit RCAT Systems Industrial UAV4 Eagle Tree Systems Flight Data Recorder Pro5 Sensors Inertial sensors 1-axis, 8 g accelerometer